Certification of Deep Learning Models for Medical Image Segmentation

TL;DR

This paper introduces a novel certification method for medical image segmentation models using randomized smoothing and diffusion models, providing theoretical guarantees and maintaining high accuracy under perturbations.

Contribution

It presents the first certified segmentation baseline for medical imaging leveraging diffusion models and randomized smoothing, establishing a foundation for future benchmarks.

Findings

High certified Dice scores on multiple datasets

Diffusion models improve certification robustness

Method maintains accuracy under significant perturbations

Abstract

In medical imaging, segmentation models have known a significant improvement in the past decade and are now used daily in clinical practice. However, similar to classification models, segmentation models are affected by adversarial attacks. In a safety-critical field like healthcare, certifying model predictions is of the utmost importance. Randomized smoothing has been introduced lately and provides a framework to certify models and obtain theoretical guarantees. In this paper, we present for the first time a certified segmentation baseline for medical imaging based on randomized smoothing and diffusion models. Our results show that leveraging the power of denoising diffusion probabilistic models helps us overcome the limits of randomized smoothing. We conduct extensive experiments on five public datasets of chest X-rays, skin lesions, and colonoscopies, and empirically show that we are able to maintain high certified Dice scores even for highly perturbed images. Our work represents the first attempt to certify medical image segmentation models, and we aspire for it to set a foundation for future benchmarks in this crucial and largely uncharted area.